Feature requests

The ability to take a vocalist you have generated in one song and have them be utilised in another. This would make it much easier to create "albums". Not entirely sure this is possible but it sure would beat the "crop/extend" casino i've been doing for nearly 90 songs now!

Advanced mode would allow sections to the prompt menu. Genre/style of song, vocalists and what sections they sing in (IE: Male lead, choir backup), instruments and such. This would also each have a "seed" so that once a proper mix was found, the seed could be copied or even saved as a default for that specific account.

First off, I love what you’ve built with Udio—it’s such a powerful tool! That said, I’ve noticed that some outputs can come out a bit low in fidelity, which got me thinking: wouldn’t it be great to have a feature that uses AI to upscale or enhance audio? Similar to how tools like Topaz Labs upscale images by filling in missing detail, this feature could intelligently add clarity and reconstruct high-frequency content in low-quality audio. It’d could also be amazing for improving recordings made at low sample rates or restoring older, degraded audio. I think this kind of feature would not only solve fidelity issues but also set Udio apart as a cutting-edge solution for audio editing. Thanks for considering, and keep up the great work!

udio-260!

Generations that are 10 to 16 seconds long, for intros or outros.

Introduce a functionality where each generated song includes separate, high-quality stems for individual elements like instruments, vocals, and effects. This would allow users to replace or modify specific components without affecting the entire track, offering complete creative freedom. By giving users the ability to customize each stem—whether swapping out a guitar for a piano or adjusting vocal tracks—this feature would provide a more versatile and dynamic music creation experience.

A slider for choosing the Degree of harshness of the vocals, for instance, when you want clean vocals on a song but it always generate screamy harsh ones.

A function for changing and for regenerating, if you need to separately replace the voice of the vocal, if you need to separately replace the instrumental itself, or the melody or rhythm or mood, singing style, genre, without changing the vocals and instrumentals, without changing the whole song. Without changing the style, vocals and instrumentals, but only its individual components and structure.

[...] People want to hear AI music that makes them laugh uncontrollably. If you take that out of the equation, you abandon what could easily be your single largest demographic. I'm honestly eager to read your opinion because it's not a simple question to answer. I'd also like to suggest you add an NSFW checkbox for people who want to mark their content that way, and for others who'd like the option to include or exclude such music from their searches. [...] From https://www.reddit.com/r/udiomusic/comments/1d2zmdd/comment/lkau64v/